Not an issue for us at the moment but we might want to think about what to do in the event that the parent job times out before the workers finish (or even start) if resources are under high demand.
Ie if a user submits a small job which doesn't complete due to workers sitting in queue for the duration, it would be good to have some transparency on that failure case.
Not an issue for us at the moment but we might want to think about what to do in the event that the parent job times out before the workers finish (or even start) if resources are under high demand.
Ie if a user submits a small job which doesn't complete due to workers sitting in queue for the duration, it would be good to have some transparency on that failure case.